1 Mean Power Output Estimation of WECs in Simulated Sea J.-B. Saulnier 1 , P. Ricci 2 , A. H. Clément 1 and A. F. de O. Falcão 3 1 Laboratoire de Mécanique des Fluides, Ecole Centrale de Nantes, 1 rue de la Noë, BP 92101, 44321 Nantes, France E-mail: jean.baptiste.saulnier@ifremer.fr, alain.clement@ec-nantes.fr 2 TECNALIA ENERGÍA, Sede de ROBOTIKER-TECNALIA, Parque tecnológico, Edificio 202, E-48170, Zamudio (Vizcaya), Spain E-mail: pricci@robotiker.es 3 IDMEC, Instituto Superior Técnico, Av. Rovisco Pais, 1, 1049-001, Lisbon, Portugal E-mail: antonio.falcao@ist.utl.pt Abstract Based on linear wave theory, two ways of simulating wave records from target spectral densities are implemented in order to assess their impact on the estimation of the mean power extracted from waves by a resonant Wave Energy Converter (WEC). The first one directly comes from the random Gaussian wave representation, while the second – widely used – derives from this latter by abusively neglecting the random nature of the individual wave amplitudes in the frequency-domain. This study investigates the consequences of such an abuse upon the device’s response estimation through the consideration of various – linear and non-linear – numerical models, and the possible improvements of the simulation time/precision compromise for further WEC design projects. Keywords: Wave signal simulation, Gaussian signals, Wave- Energy Converter, Mean power estimation. Nomenclature a i ,b i ,u i = Fourier series frequency amplitudes C PTO = PTO damping coefficient E = variance spectral density f,ω = wave frequency f Nyq = Nyquist frequency H m0 = significant wave height (= 4m 0 1/2 ) m 0 = 0 th -order spectral moment, variance N t = number of time-simulation points P PTO = absorbed power by PTO © Proceedings of the 8th European Wave and Tidal Energy Conference, Uppsala, Sweden, 2009 R(τ) = auto-correlation function t = time T = time-simulation length z = heave motion η = sea-surface elevation ϕ = wave phase σ = standard deviation ∆f = frequency bin ∆t = simulation time-step Λ = equivalent spectral bandwidth Subscripts i = frequency harmonic i T = time-simulation over [0;T] Superscripts - , < > = mean value (time, ensemble) ^ = parameter estimator i = time-series discrete point or ensemble element i . = time first derivative 1 Introduction Full-scale devices or prototypes monitoring data are still rare to obtain although new near-shore projects arose in the last few years installed some small pilot units at sea (OEBuoy off the Galway Bay, Pelamis at Aguçadoura…). Numerical simulation models still are – and will be – necessary to predict and/or complement the field data, namely prior to any in situ deployment. In most of these models, linear theory is invoked to represent the undisturbed wave-field from which energy is absorbed by the device. From spectral densities (omnidirectional or directional) random time- 891